Comparison of Multiple Bioactive Constituents in Different Parts of Eucommia ulmoides Based on UFLC-QTRAP-MS/MS Combined with PCA

Eucommia ulmoides Oilv. (EU), also called Du-zhong, is a classical traditional Chinese medicine. Its bark, leaf, and male flower are all used for medicinal purposes, called Eucommiae Cortex (EC), Eucommiae Folium (EF), and Eucommiae Flos Male (EFM). In order to study the difference in synthesis and the accumulation of metabolites in different parts of EU, a reliable method based on ultra-fast liquid chromatography tandem triple quadrupole mass spectrometry (UFLC-QTRAP-MS/MS) was developed for the simultaneous determination of a total of 21 constituents, including two lignans, 6 iridoids, 6 penylpropanoids, 6 flavonoids, and one phenol in the samples (EC, EF, and EFM). Furthermore, principal component analysis (PCA) was performed to evaluate and classify the samples according to the contents of these 21 constituents. All of the results demonstrated that the chemical compositions in EC, EF, and EFM were significantly different and the differential constituents (i.e., aucubin, geniposidic acid, chlorogenic acid, pinoresinol-di-O-β-d-glucopyranoside, geniposide, cryptochlorogenic acid, rutin, and quercetin) were remarkably associated with sample classifications. The research will provide the basic information for revealing the laws of metabolite accumulation in EC, EF, and EFM from the same origin.


Introduction
Eucommiae Cortex (EC), Eucommiae Folium (EF), and Eucommiae Flos Male (EFM) are derived from the dried bark, leaf, and male flower of Eucommia ulmoides Oilv. (EU), respectively. EC has been used as an important traditional Chinese medicine (TCM) for more than 2000 years in China, and EF has become a popular functional health food and plant medicine material in past twenty years. EC and EF are officially documented in the Chinese Pharmacopoeia [1]. Though the studies about EFM started relatively late, accumulating evidence has proved that EFM, like EC and EF, was found to be rich in bioactive constituents [2][3][4].
Phytochemical investigations have revealed that EU mainly contains several types of constituents, such as lignans, iridoids, penylpropanoids, and flavonoids [5]. Pharmacological studies demonstrated that lignans have biological activities including anti-hypertensive, anti-tumor, anti-inflammatory, liver-protection, and inhibiting platelet activation; iridoids are known to have a variety of biological activities such as improvement in the collagen synthesis, anti-aging, anti-tumor, anti-obesity

Plant Materials
The EC (S1-S3), EF (S4-S6), and EFM (S7-S9) samples were collected from Lueyang City, Shaanxi Province (105 • 42 53" N, 33 • 23 6" E) in the traditional harvest time and dried at source area. The samples were authenticated by Prof. Xunhong Liu of the Nanjing University of Chinese Medicine, and the voucher specimens were deposited at the Herbarium in School of Pharmacy, Nanjing University of Chinese Medicine, China.

Preparation of Standard Solution
A mixed standard stock solution containing 21 reference substances was prepared by dissolving them in methanol and their concentrations were as follows: 1, 7

Preparation of Sample Solutions
0.6 g of sample powder, after passing through a 40 mesh sieve, was weighed accurately and ultrasonically extracted with 30 mL 50% (v/v) methanol for 20 min and cooled at room temperature, then 50% (v/v) methanol was added to compensate for the lost weight. The resultant solution was subsequently centrifuged at 12,000 rpm for 10 min and the supernatant was stored at 4 • C and filtered through a 0.22 µm membrane (Jinteng laboratory equipment Co., Ltd., Tianjin, China) prior to injection for LC-MS analysis.
Mass spectrometry was performed using an API 5500 triple quadrupole mass spectrometer (AB Sciex, Framingham, MA, USA) equipped with an electrospray ionization (ESI) source operating in both positive and negative ion modes. The parameters in the source were set as follows: GS1 flow, 65 L/min; GS2 flow, 65 L/min; CUR flow, 30 L/min; gas temperature, 650 • C; pressures of nebulizer of MS, 4500 V (positive) and −4500 V (negative). All MS data were acquired using the Analyst 1.6.2 software to ensure mass accuracy and reproducibility.

Validation of The Method
The proposed method was validated according to the International Conference on Harmonisation (ICH) guidelines Q2 (R1) [21]. The principal parameters studied were linearity and range, limits of detection and quantification (LOD and LOQ), precision, repeatability, solution stability, and accuracy. The standard solution containing 21 compounds was prepared and diluted with methanol to appropriate concentrations for the construction of calibration curves. Calibration curves were developed by plotting the peak areas versus the corresponding concentrations of each analyte. The correlation coefficient, slope, and y-intercept were calculated for each replicate with acceptance criteria of a correlation coefficient r ≥ 0.99.
The LOD and LOQ of 21 compounds were measured at signal-to-noise ratios of 3 and 10, respectively.

Precision, Repeatability, Solution Stability, and Accuracy
Intra-and inter-day variations were used to evaluate the precision of the established method. The relative standard deviation (RSD) of the peak area was taken as a measure of precision.
To confirm the repeatability, six different analytical sample solutions prepared from the same sample were analyzed. The RSD of the peak area was taken as a measure of repeatability.
A stability test was further performed to analyze the variations in the sample solutions at 0, 2, 4, 8, 12, and 24 h, respectively.
A recovery test was used to evaluate the accuracy of the established method. The test was performed by adding the corresponding marker constituents at low (80% of the known amounts), medium (same as the known amounts), and high (120% of the known amounts) levels to the EC sample which had been analyzed previously. The mixture was extracted and analyzed using the aforementioned method in triplicate.

Principal Component Analysis (PCA)
PCA is an unsupervised pattern recognition method used for analyzing, classifying, and reducing the dimensionality of numerical datasets in a multivariate problem [22], and it has been widely used for the quality control of herbal medicines [23][24][25]. Data of the contents of 21 compounds in EC, EF, and EFM samples were listed in an Excel file. PCA was used to evaluate the variations of the three different parts of Eucommia ulmoides according to the contents of the 21 bioactive compounds analyzed in this study using Simca-P 13.0 (version 13.0, Umetrics AB, Umea, Sweden) software. The data which shown in Table 3 was imported into Simca-P 13.0 and centralized by Simca-P 13.0, and a data matrix after centralization is obtained which was shown in Table S1. When the contents of investigated compounds were below the quantitation limit or not detected in the samples, the values of such elements were considered to be 0.

Optimization of Extraction Conditions
In order to achieve an efficient extraction of bioactive constituents in samples, orthogonal test was employed to investigate these key factors, including methanol concentration (30% (v/v) methanol, 50% (v/v) methanol, and 80% (v/v) methanol), sample-solvent ratio (1:30, 1:50, and 1:80 (w/v)), and ultrasonic time (10 min, 20 min, and 40 min). Finally, the optimum sample extraction condition was achieved by ultrasonic extraction with a 1:50 ratio of 50% methanol for 20 min. All of the samples were extracted at room temperature.

Optimization of MS Conditions
Preliminary experiments were conducted with the purpose of finding the best instrumental conditions. The individual solutions of all standard compounds (100 ng/mL in methanol) were injected into the ESI source in the positive and negative ion modes. After trial and error inspection, other compounds have a good condition in the negative ion mode, while syringin has a good condition in the positive ion mode. Most abundance fragment ions were selected as MRM transition from MS/MS spectrum, and the highest sensitivity was obtained at a certain value of fragmentor and collision energy (CE). The optimum values for each condition of 21 compounds were summarized in Table 1.

Linearity and Range, LOD, and LOQ
The calibration curve for each compound was obtained in duplicate with at least six appropriate concentrations. The correlation coefficients of all target components exceeded 0.9991 with good linearity. The LODs and LOQs of 21 compounds were measured at signal-to-noise ratios of 3 and 10 and the ranges were 0.32-4.76 ng/mL and 1.15-16.66 ng/mL, respectively. The results were shown in Table 2.

Precision, Repeatability, Solution Stability, and Accuracy
For the intra-day variability test, the mixed standard solutions were analyzed for six replicates within a day; for the inter-day variability test, the solutions were examined for three consecutive days. The relative standard deviation (RSD) was taken as a measure of precision. The RSD values of intra-and inter-day variations of 21 compounds were in the range of 1.09%-3.45% and 2.10%-3.76%, respectively. The results were shown in Table 2.
The satisfactory repeatabilities presented as RSD values were in the range of 1.02% to 3.68%. The results were shown in Table 2.
The solution stabilities presented as RSD values were less than 3.85%, indicating the sample solution was stable when stored at room temperature for 24 h. The results were shown in Table 2.
The overall recoveries lay between 95.83% and 103.30%, with RSD values between 1.03% and 3.98%. The results were shown in Table 2.

Sample Analysis
The developed UFLC-QTRAP-MS/MS method was subsequently applied to the comprehensive quality evaluation of EC, EF, and EFM samples. The contents of 21 compounds were simultaneously determinated in this study. In 10 metabolites isolated from EU by the authors, which were determinated in the study by Chai et al. [17], we determinated 8 metabolites. The contents of wogonin and oroxylin A in the samples were too low to meet the linear range. Additionally, the standards of licoagroside F and syringaresinol di-O-β-D-glucopyranoside were not obtained. Typical MRM chromatograms were shown in Figure 1, and the results of the quantitative determination of 21 compounds from these samples were summarized in Table 3. The results were reported as mean ± SD.
By comparing the amounts, it was found that the constituents of EC, EF, and EFM samples were quite different. The total contents of 21 constituents varied from 5132.65 µg/g to 34,200.10 µg/g. The total contents of each type of constituent were also calculated; two lignans ranged from 16.37 µg/g to 1497.00 µg/g, 6 iridoids ranged from 172.48 µg/g to 10,187.34 µg/g, 6 penylpropanoids ranged from 2128.32 µg/g to 21,421.96 µg/g, 6 flavonoids ranged from 0.21 µg/g to 4944.46 µg/g, and one phenol ranged from 4.85 µg/g to 262.80 µg/g. The results indicated that the contents of the 21 compounds were obviously different in the three different parts of EC.

Conclusions
In this study, an efficient and sensitive UFLC-QTRAP-MS/MS method has been developed and validated for the simultaneous determination of a total of 21 constituents, including two lignans, 6 iridoids, 6 penylpropanoids, 6 flavonoids, and one phenol in EU samples. The validated method was successfully applied to quantify 21 bioactive constituents in EC, EF, and EFM. PCA was performed to evaluate and classify the samples according to the contents of these 21 constituents. All of the results demonstrated that the chemical compositions in EC, EF, and EFM were significantly different, and the differential constituents (i.e., aucubin, geniposidic acid, chlorogenic acid, pinoresinol-di-O-β-D-glucopyranoside, geniposide, cryptochlorogenic acid, rutin, and quercetin) had a significant relationship with the sample classifications. This research will provide the basic foundation for revealing the laws of metabolite accumulation in EC, EF, and EFM from the same origin.
Supplementary Materials: The following are available online. Figure S1: Chemical structures of 21 reference substances, Table S1: The data matrix applying PCA, Table S2: The p1 and p2 values of the points in PCA loading plot.